1 /* handy.h 2 * 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000, 4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others 5 * 6 * You may distribute under the terms of either the GNU General Public 7 * License or the Artistic License, as specified in the README file. 8 * 9 */ 10 11 /* IMPORTANT NOTE: Everything whose name begins with an underscore is for 12 * internal core Perl use only. */ 13 14 #ifndef HANDY_H /* Guard against nested #inclusion */ 15 #define HANDY_H 16 17 #if !defined(__STDC__) 18 #ifdef NULL 19 #undef NULL 20 #endif 21 # define NULL 0 22 #endif 23 24 #ifndef PERL_CORE 25 # define Null(type) ((type)NULL) 26 27 /* 28 =head1 Handy Values 29 30 =for apidoc AmU||Nullch 31 Null character pointer. (No longer available when C<PERL_CORE> is 32 defined.) 33 34 =for apidoc AmU||Nullsv 35 Null SV pointer. (No longer available when C<PERL_CORE> is defined.) 36 37 =cut 38 */ 39 40 # define Nullch Null(char*) 41 # define Nullfp Null(PerlIO*) 42 # define Nullsv Null(SV*) 43 #endif 44 45 #ifdef TRUE 46 #undef TRUE 47 #endif 48 #ifdef FALSE 49 #undef FALSE 50 #endif 51 #define TRUE (1) 52 #define FALSE (0) 53 54 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way 55 * (compiler permitting) that casting away const-ness will give a warning; 56 * e.g.: 57 * 58 * const SV *sv = ...; 59 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away 60 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn 61 */ 62 63 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) 64 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; }) 65 #else 66 # define MUTABLE_PTR(p) ((void *) (p)) 67 #endif 68 69 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p)) 70 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p)) 71 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p)) 72 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p)) 73 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p)) 74 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p)) 75 76 #if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR) 77 # include <stdbool.h> 78 # ifndef HAS_BOOL 79 # define HAS_BOOL 1 80 # endif 81 #endif 82 83 /* bool is built-in for g++-2.6.3 and later, which might be used 84 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't 85 be sure _G_config.h will be included before this file. _G_config.h 86 also defines _G_HAVE_BOOL for both gcc and g++, but only g++ 87 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us. 88 g++ can be identified by __GNUG__. 89 Andy Dougherty February 2000 90 */ 91 #ifdef __GNUG__ /* GNU g++ has bool built-in */ 92 # ifndef PERL_BOOL_AS_CHAR 93 # ifndef HAS_BOOL 94 # define HAS_BOOL 1 95 # endif 96 # endif 97 #endif 98 99 #ifndef HAS_BOOL 100 # ifdef bool 101 # undef bool 102 # endif 103 # define bool char 104 # define HAS_BOOL 1 105 #endif 106 107 /* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is 108 * defined as char for example, then the cast from int is 109 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on 110 * AIX */ 111 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0) 112 113 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any. 114 * XXX Should really be a Configure probe, with HAS__FUNCTION__ 115 * and FUNCTION__ as results. 116 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */ 117 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */ 118 # define FUNCTION__ __func__ 119 #else 120 # if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \ 121 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */ 122 # define FUNCTION__ "" 123 # else 124 # define FUNCTION__ __FUNCTION__ /* Common extension. */ 125 # endif 126 #endif 127 128 /* XXX A note on the perl source internal type system. The 129 original intent was that I32 be *exactly* 32 bits. 130 131 Currently, we only guarantee that I32 is *at least* 32 bits. 132 Specifically, if int is 64 bits, then so is I32. (This is the case 133 for the Cray.) This has the advantage of meshing nicely with 134 standard library calls (where we pass an I32 and the library is 135 expecting an int), but the disadvantage that an I32 is not 32 bits. 136 Andy Dougherty August 1996 137 138 There is no guarantee that there is *any* integral type with 139 exactly 32 bits. It is perfectly legal for a system to have 140 sizeof(short) == sizeof(int) == sizeof(long) == 8. 141 142 Similarly, there is no guarantee that I16 and U16 have exactly 16 143 bits. 144 145 For dealing with issues that may arise from various 32/64-bit 146 systems, we will ask Configure to check out 147 148 SHORTSIZE == sizeof(short) 149 INTSIZE == sizeof(int) 150 LONGSIZE == sizeof(long) 151 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG) 152 PTRSIZE == sizeof(void *) 153 DOUBLESIZE == sizeof(double) 154 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE). 155 156 */ 157 158 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */ 159 # include <inttypes.h> 160 # ifdef INT32_MIN_BROKEN 161 # undef INT32_MIN 162 # define INT32_MIN (-2147483647-1) 163 # endif 164 # ifdef INT64_MIN_BROKEN 165 # undef INT64_MIN 166 # define INT64_MIN (-9223372036854775807LL-1) 167 # endif 168 #endif 169 170 typedef I8TYPE I8; 171 typedef U8TYPE U8; 172 typedef I16TYPE I16; 173 typedef U16TYPE U16; 174 typedef I32TYPE I32; 175 typedef U32TYPE U32; 176 177 #ifdef HAS_QUAD 178 typedef I64TYPE I64; 179 typedef U64TYPE U64; 180 #endif 181 182 /* INT64_C/UINT64_C are C99 from <stdint.h> (so they will not be 183 * available in strict C89 mode), but they are nice, so let's define 184 * them if necessary. */ 185 #if defined(HAS_QUAD) 186 # undef PeRl_INT64_C 187 # undef PeRl_UINT64_C 188 /* Prefer the native integer types (int and long) over long long 189 * (which is not C89) and Win32-specific __int64. */ 190 # if QUADKIND == QUAD_IS_INT && INTSIZE == 8 191 # define PeRl_INT64_C(c) (c) 192 # define PeRl_UINT64_C(c) CAT2(c,U) 193 # endif 194 # if QUADKIND == QUAD_IS_LONG && LONGSIZE == 8 195 # define PeRl_INT64_C(c) CAT2(c,L) 196 # define PeRl_UINT64_C(c) CAT2(c,UL) 197 # endif 198 # if QUADKIND == QUAD_IS_LONG_LONG && defined(HAS_LONG_LONG) 199 # define PeRl_INT64_C(c) CAT2(c,LL) 200 # define PeRl_UINT64_C(c) CAT2(c,ULL) 201 # endif 202 # if QUADKIND == QUAD_IS___INT64 203 # define PeRl_INT64_C(c) CAT2(c,I64) 204 # define PeRl_UINT64_C(c) CAT2(c,UI64) 205 # endif 206 # ifndef PeRl_INT64_C 207 # define PeRl_INT64_C(c) ((I64)(c)) /* last resort */ 208 # define PeRl_UINT64_C(c) ((U64)(c)) 209 # endif 210 /* In OS X the INT64_C/UINT64_C are defined with LL/ULL, which will 211 * not fly with C89-pedantic gcc, so let's undefine them first so that 212 * we can redefine them with our native integer preferring versions. */ 213 # if defined(PERL_DARWIN) && defined(PERL_GCC_PEDANTIC) 214 # undef INT64_C 215 # undef UINT64_C 216 # endif 217 # ifndef INT64_C 218 # define INT64_C(c) PeRl_INT64_C(c) 219 # endif 220 # ifndef UINT64_C 221 # define UINT64_C(c) PeRl_UINT64_C(c) 222 # endif 223 #endif 224 225 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX) 226 227 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. 228 Please search CHAR_MAX in perl.h for further details. */ 229 #define U8_MAX UINT8_MAX 230 #define U8_MIN UINT8_MIN 231 232 #define I16_MAX INT16_MAX 233 #define I16_MIN INT16_MIN 234 #define U16_MAX UINT16_MAX 235 #define U16_MIN UINT16_MIN 236 237 #define I32_MAX INT32_MAX 238 #define I32_MIN INT32_MIN 239 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */ 240 # define U32_MAX UINT32_MAX 241 #else 242 # define U32_MAX 4294967295U 243 #endif 244 #define U32_MIN UINT32_MIN 245 246 #else 247 248 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. 249 Please search CHAR_MAX in perl.h for further details. */ 250 #define U8_MAX PERL_UCHAR_MAX 251 #define U8_MIN PERL_UCHAR_MIN 252 253 #define I16_MAX PERL_SHORT_MAX 254 #define I16_MIN PERL_SHORT_MIN 255 #define U16_MAX PERL_USHORT_MAX 256 #define U16_MIN PERL_USHORT_MIN 257 258 #if LONGSIZE > 4 259 # define I32_MAX PERL_INT_MAX 260 # define I32_MIN PERL_INT_MIN 261 # define U32_MAX PERL_UINT_MAX 262 # define U32_MIN PERL_UINT_MIN 263 #else 264 # define I32_MAX PERL_LONG_MAX 265 # define I32_MIN PERL_LONG_MIN 266 # define U32_MAX PERL_ULONG_MAX 267 # define U32_MIN PERL_ULONG_MIN 268 #endif 269 270 #endif 271 272 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */ 273 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */ 274 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8) 275 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */ 276 277 /* Unused by core; should be deprecated */ 278 #define Ctl(ch) ((ch) & 037) 279 280 /* This is a helper macro to avoid preprocessor issues, replaced by nothing 281 * unless under DEBUGGING, where it expands to an assert of its argument, 282 * followed by a comma (hence the comma operator). If we just used a straight 283 * assert(), we would get a comma with nothing before it when not DEBUGGING. 284 * 285 * We also use empty definition under Coverity since the __ASSERT__ 286 * checks often check for things that Really Cannot Happen, and Coverity 287 * detects that and gets all excited. */ 288 289 #if defined(DEBUGGING) && !defined(__COVERITY__) 290 # define __ASSERT_(statement) assert(statement), 291 #else 292 # define __ASSERT_(statement) 293 #endif 294 295 /* 296 =head1 SV-Body Allocation 297 298 =for apidoc Ama|SV*|newSVpvs|const char* s 299 Like C<newSVpvn>, but takes a C<NUL>-terminated literal string instead of a 300 string/length pair. 301 302 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags 303 Like C<newSVpvn_flags>, but takes a C<NUL>-terminated literal string instead of 304 a string/length pair. 305 306 =for apidoc Ama|SV*|newSVpvs_share|const char* s 307 Like C<newSVpvn_share>, but takes a C<NUL>-terminated literal string instead of 308 a string/length pair and omits the hash parameter. 309 310 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags 311 Like C<sv_catpvn_flags>, but takes a C<NUL>-terminated literal string instead 312 of a string/length pair. 313 314 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s 315 Like C<sv_catpvn_nomg>, but takes a C<NUL>-terminated literal string instead of 316 a string/length pair. 317 318 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s 319 Like C<sv_catpvn>, but takes a C<NUL>-terminated literal string instead of a 320 string/length pair. 321 322 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s 323 Like C<sv_catpvn_mg>, but takes a C<NUL>-terminated literal string instead of a 324 string/length pair. 325 326 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s 327 Like C<sv_setpvn>, but takes a C<NUL>-terminated literal string instead of a 328 string/length pair. 329 330 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s 331 Like C<sv_setpvn_mg>, but takes a C<NUL>-terminated literal string instead of a 332 string/length pair. 333 334 =for apidoc Am|SV *|sv_setref_pvs|const char* s 335 Like C<sv_setref_pvn>, but takes a C<NUL>-terminated literal string instead of 336 a string/length pair. 337 338 =head1 Memory Management 339 340 =for apidoc Ama|char*|savepvs|const char* s 341 Like C<savepvn>, but takes a C<NUL>-terminated literal string instead of a 342 string/length pair. 343 344 =for apidoc Ama|char*|savesharedpvs|const char* s 345 A version of C<savepvs()> which allocates the duplicate string in memory 346 which is shared between threads. 347 348 =head1 GV Functions 349 350 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create 351 Like C<gv_stashpvn>, but takes a C<NUL>-terminated literal string instead of a 352 string/length pair. 353 354 =head1 Hash Manipulation Functions 355 356 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval 357 Like C<hv_fetch>, but takes a C<NUL>-terminated literal string instead of a 358 string/length pair. 359 360 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val 361 Like C<hv_store>, but takes a C<NUL>-terminated literal string instead of a 362 string/length pair 363 and omits the hash parameter. 364 365 =head1 Lexer interface 366 367 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags 368 369 Like L</lex_stuff_pvn>, but takes a C<NUL>-terminated literal string instead of 370 a string/length pair. 371 372 =cut 373 */ 374 375 /* concatenating with "" ensures that only literal strings are accepted as 376 * argument */ 377 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1) 378 379 /* note that STR_WITH_LEN() can't be used as argument to macros or functions 380 * that under some configurations might be macros, which means that it requires 381 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used. 382 */ 383 384 /* STR_WITH_LEN() shortcuts */ 385 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str)) 386 #define newSVpvs_flags(str,flags) \ 387 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags) 388 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0) 389 #define sv_catpvs_flags(sv, str, flags) \ 390 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags) 391 #define sv_catpvs_nomg(sv, str) \ 392 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0) 393 #define sv_catpvs(sv, str) \ 394 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC) 395 #define sv_catpvs_mg(sv, str) \ 396 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC) 397 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str)) 398 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str)) 399 #define sv_setref_pvs(rv, classname, str) \ 400 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str)) 401 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str)) 402 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str)) 403 #define gv_stashpvs(str, create) \ 404 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create) 405 #define gv_fetchpvs(namebeg, add, sv_type) \ 406 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type) 407 #define gv_fetchpvn(namebeg, len, add, sv_type) \ 408 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type) 409 #define sv_catxmlpvs(dsv, str, utf8) \ 410 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8) 411 #define hv_fetchs(hv,key,lval) \ 412 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \ 413 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \ 414 : HV_FETCH_JUST_SV, NULL, 0)) 415 416 #define hv_stores(hv,key,val) \ 417 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \ 418 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0)) 419 420 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags) 421 422 #define get_cvs(str, flags) \ 423 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags)) 424 425 /* 426 =head1 Miscellaneous Functions 427 428 =for apidoc Am|bool|strNE|char* s1|char* s2 429 Test two C<NUL>-terminated strings to see if they are different. Returns true 430 or false. 431 432 =for apidoc Am|bool|strEQ|char* s1|char* s2 433 Test two C<NUL>-terminated strings to see if they are equal. Returns true or 434 false. 435 436 =for apidoc Am|bool|strLT|char* s1|char* s2 437 Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than the 438 second, C<s2>. Returns true or false. 439 440 =for apidoc Am|bool|strLE|char* s1|char* s2 441 Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than or 442 equal to the second, C<s2>. Returns true or false. 443 444 =for apidoc Am|bool|strGT|char* s1|char* s2 445 Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than 446 the second, C<s2>. Returns true or false. 447 448 =for apidoc Am|bool|strGE|char* s1|char* s2 449 Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than 450 or equal to the second, C<s2>. Returns true or false. 451 452 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len 453 Test two C<NUL>-terminated strings to see if they are different. The C<len> 454 parameter indicates the number of bytes to compare. Returns true or false. (A 455 wrapper for C<strncmp>). 456 457 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len 458 Test two C<NUL>-terminated strings to see if they are equal. The C<len> 459 parameter indicates the number of bytes to compare. Returns true or false. (A 460 wrapper for C<strncmp>). 461 462 =for apidoc Am|bool|memEQ|char* s1|char* s2|STRLEN len 463 Test two buffers (which may contain embedded C<NUL> characters, to see if they 464 are equal. The C<len> parameter indicates the number of bytes to compare. 465 Returns zero if equal, or non-zero if non-equal. 466 467 =for apidoc Am|bool|memNE|char* s1|char* s2|STRLEN len 468 Test two buffers (which may contain embedded C<NUL> characters, to see if they 469 are not equal. The C<len> parameter indicates the number of bytes to compare. 470 Returns zero if non-equal, or non-zero if equal. 471 472 =cut 473 */ 474 475 #define strNE(s1,s2) (strcmp(s1,s2)) 476 #define strEQ(s1,s2) (!strcmp(s1,s2)) 477 #define strLT(s1,s2) (strcmp(s1,s2) < 0) 478 #define strLE(s1,s2) (strcmp(s1,s2) <= 0) 479 #define strGT(s1,s2) (strcmp(s1,s2) > 0) 480 #define strGE(s1,s2) (strcmp(s1,s2) >= 0) 481 #define strnNE(s1,s2,l) (strncmp(s1,s2,l)) 482 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l)) 483 484 #ifdef HAS_MEMCMP 485 # define memNE(s1,s2,l) (memcmp(s1,s2,l)) 486 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l)) 487 #else 488 # define memNE(s1,s2,l) (bcmp(s1,s2,l)) 489 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l)) 490 #endif 491 492 #define memEQs(s1, l, s2) \ 493 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1))) 494 #define memNEs(s1, l, s2) !memEQs(s1, l, s2) 495 496 /* 497 * Character classes. 498 * 499 * Unfortunately, the introduction of locales means that we 500 * can't trust isupper(), etc. to tell the truth. And when 501 * it comes to /\w+/ with tainting enabled, we *must* be able 502 * to trust our character classes. 503 * 504 * Therefore, the default tests in the text of Perl will be 505 * independent of locale. Any code that wants to depend on 506 * the current locale will use the tests that begin with "lc". 507 */ 508 509 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */ 510 # ifndef CTYPE256 511 # define CTYPE256 512 # endif 513 #endif 514 515 /* 516 517 =head1 Character classification 518 This section is about functions (really macros) that classify characters 519 into types, such as punctuation versus alphabetic, etc. Most of these are 520 analogous to regular expression character classes. (See 521 L<perlrecharclass/POSIX Character Classes>.) There are several variants for 522 each class. (Not all macros have all variants; each item below lists the 523 ones valid for it.) None are affected by C<use bytes>, and only the ones 524 with C<LC> in the name are affected by the current locale. 525 526 The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a 527 C<U8>) as input and returns a boolean as to whether or not the character 528 represented by that octet is (or on non-ASCII platforms, corresponds to) an 529 ASCII character in the named class based on platform, Unicode, and Perl rules. 530 If the input is a number that doesn't fit in an octet, FALSE is returned. 531 532 Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function 533 with no suffix C<"_A">. 534 535 Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set 536 onto the platform. That is, the code points that are ASCII are unaffected, 537 since ASCII is a subset of Latin-1. But the non-ASCII code points are treated 538 as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return 539 true when called with the code point 0xDF, which is a word character in both 540 ASCII and EBCDIC (though it represents different characters in each). 541 542 Variant C<isFOO_uvchr> is like the C<isFOO_L1> variant, but accepts any UV code 543 point as input. If the code point is larger than 255, Unicode rules are used 544 to determine if it is in the character class. For example, 545 C<isWORDCHAR_uvchr(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A 546 WITH MACRON in Unicode, and is a word character. 547 548 Variant C<isFOO_utf8> is like C<isFOO_uvchr>, but the input is a pointer to a 549 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>, and 550 possibly containing embedded C<NUL> characters). The classification of just 551 the first (possibly multi-byte) character in the string is tested. 552 553 Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the 554 result is based on the current locale, which is what C<LC> in the name stands 555 for. If Perl can determine that the current locale is a UTF-8 locale, it uses 556 the published Unicode rules; otherwise, it uses the C library function that 557 gives the named classification. For example, C<isDIGIT_LC()> when not in a 558 UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always 559 returned if the input won't fit into an octet. On some platforms where the C 560 library function is known to be defective, Perl changes its result to follow 561 the POSIX standard's rules. 562 563 Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It 564 returns the same as C<isFOO_LC> for input code points less than 256, and 565 returns the hard-coded, not-affected-by-locale, Unicode results for larger ones. 566 567 Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer 568 to a (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>, and 569 possibly containing embedded C<NUL> characters). The classification of just 570 the first (possibly multi-byte) character in the string is tested. 571 572 =for apidoc Am|bool|isALPHA|char ch 573 Returns a boolean indicating whether the specified character is an 574 alphabetic character, analogous to C<m/[[:alpha:]]/>. 575 See the L<top of this section|/Character classification> for an explanation of 576 variants 577 C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uvchr>, C<isALPHA_utf8>, C<isALPHA_LC>, 578 C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>. 579 580 =for apidoc Am|bool|isALPHANUMERIC|char ch 581 Returns a boolean indicating whether the specified character is a either an 582 alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>. 583 See the L<top of this section|/Character classification> for an explanation of 584 variants 585 C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uvchr>, 586 C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and 587 C<isALPHANUMERIC_LC_utf8>. 588 589 =for apidoc Am|bool|isASCII|char ch 590 Returns a boolean indicating whether the specified character is one of the 128 591 characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>. 592 On non-ASCII platforms, it returns TRUE iff this 593 character corresponds to an ASCII character. Variants C<isASCII_A()> and 594 C<isASCII_L1()> are identical to C<isASCII()>. 595 See the L<top of this section|/Character classification> for an explanation of 596 variants 597 C<isASCII_uvchr>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and 598 C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C 599 library routine C<isascii()>. In these cases, the variants whose names contain 600 C<LC> are the same as the corresponding ones without. 601 602 Also note, that because all ASCII characters are UTF-8 invariant (meaning they 603 have the exact same representation (always a single byte) whether encoded in 604 UTF-8 or not), C<isASCII> will give the correct results when called with any 605 byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will 606 work properly on any string encoded or not in UTF-8. 607 608 =for apidoc Am|bool|isBLANK|char ch 609 Returns a boolean indicating whether the specified character is a 610 character considered to be a blank, analogous to C<m/[[:blank:]]/>. 611 See the L<top of this section|/Character classification> for an explanation of 612 variants 613 C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uvchr>, C<isBLANK_utf8>, C<isBLANK_LC>, 614 C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some 615 platforms do not have the C library routine C<isblank()>. In these cases, the 616 variants whose names contain C<LC> are the same as the corresponding ones 617 without. 618 619 =for apidoc Am|bool|isCNTRL|char ch 620 Returns a boolean indicating whether the specified character is a 621 control character, analogous to C<m/[[:cntrl:]]/>. 622 See the L<top of this section|/Character classification> for an explanation of 623 variants 624 C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uvchr>, C<isCNTRL_utf8>, C<isCNTRL_LC>, 625 C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8> 626 On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant. 627 628 =for apidoc Am|bool|isDIGIT|char ch 629 Returns a boolean indicating whether the specified character is a 630 digit, analogous to C<m/[[:digit:]]/>. 631 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>. 632 See the L<top of this section|/Character classification> for an explanation of 633 variants 634 C<isDIGIT_uvchr>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and 635 C<isDIGIT_LC_utf8>. 636 637 =for apidoc Am|bool|isGRAPH|char ch 638 Returns a boolean indicating whether the specified character is a 639 graphic character, analogous to C<m/[[:graph:]]/>. 640 See the L<top of this section|/Character classification> for an explanation of 641 variants 642 C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uvchr>, C<isGRAPH_utf8>, C<isGRAPH_LC>, 643 C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>. 644 645 =for apidoc Am|bool|isLOWER|char ch 646 Returns a boolean indicating whether the specified character is a 647 lowercase character, analogous to C<m/[[:lower:]]/>. 648 See the L<top of this section|/Character classification> for an explanation of 649 variants 650 C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uvchr>, C<isLOWER_utf8>, C<isLOWER_LC>, 651 C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>. 652 653 =for apidoc Am|bool|isOCTAL|char ch 654 Returns a boolean indicating whether the specified character is an 655 octal digit, [0-7]. 656 The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to 657 C<isOCTAL>. 658 659 =for apidoc Am|bool|isPUNCT|char ch 660 Returns a boolean indicating whether the specified character is a 661 punctuation character, analogous to C<m/[[:punct:]]/>. 662 Note that the definition of what is punctuation isn't as 663 straightforward as one might desire. See L<perlrecharclass/POSIX Character 664 Classes> for details. 665 See the L<top of this section|/Character classification> for an explanation of 666 variants 667 C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uvchr>, C<isPUNCT_utf8>, C<isPUNCT_LC>, 668 C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>. 669 670 =for apidoc Am|bool|isSPACE|char ch 671 Returns a boolean indicating whether the specified character is a 672 whitespace character. This is analogous 673 to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18 674 this also matches what C<m/[[:space:]]/> does. Prior to 5.18, only the 675 locale forms of this macro (the ones with C<LC> in their names) matched 676 precisely what C<m/[[:space:]]/> does. In those releases, the only difference, 677 in the non-locale variants, was that C<isSPACE()> did not match a vertical tab. 678 (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.) 679 See the L<top of this section|/Character classification> for an explanation of 680 variants 681 C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uvchr>, C<isSPACE_utf8>, C<isSPACE_LC>, 682 C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>. 683 684 =for apidoc Am|bool|isPSXSPC|char ch 685 (short for Posix Space) 686 Starting in 5.18, this is identical in all its forms to the 687 corresponding C<isSPACE()> macros. 688 The locale forms of this macro are identical to their corresponding 689 C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the 690 non-locale forms differ from their C<isSPACE()> forms only in that the 691 C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do. 692 Otherwise they are identical. Thus this macro is analogous to what 693 C<m/[[:space:]]/> matches in a regular expression. 694 See the L<top of this section|/Character classification> for an explanation of 695 variants C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uvchr>, C<isPSXSPC_utf8>, 696 C<isPSXSPC_LC>, C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>. 697 698 =for apidoc Am|bool|isUPPER|char ch 699 Returns a boolean indicating whether the specified character is an 700 uppercase character, analogous to C<m/[[:upper:]]/>. 701 See the L<top of this section|/Character classification> for an explanation of 702 variants 703 C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uvchr>, C<isUPPER_utf8>, C<isUPPER_LC>, 704 C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>. 705 706 =for apidoc Am|bool|isPRINT|char ch 707 Returns a boolean indicating whether the specified character is a 708 printable character, analogous to C<m/[[:print:]]/>. 709 See the L<top of this section|/Character classification> for an explanation of 710 variants 711 C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uvchr>, C<isPRINT_utf8>, C<isPRINT_LC>, 712 C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>. 713 714 =for apidoc Am|bool|isWORDCHAR|char ch 715 Returns a boolean indicating whether the specified character is a character 716 that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match 717 in a regular expression. A word character is an alphabetic character, a 718 decimal digit, a connecting punctuation character (such as an underscore), or 719 a "mark" character that attaches to one of those (like some sort of accent). 720 C<isALNUM()> is a synonym provided for backward compatibility, even though a 721 word character includes more than the standard C language meaning of 722 alphanumeric. 723 See the L<top of this section|/Character classification> for an explanation of 724 variants 725 C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uvchr>, and C<isWORDCHAR_utf8>. 726 C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8> are also as 727 described there, but additionally include the platform's native underscore. 728 729 =for apidoc Am|bool|isXDIGIT|char ch 730 Returns a boolean indicating whether the specified character is a hexadecimal 731 digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()> 732 and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>. 733 See the L<top of this section|/Character classification> for an explanation of 734 variants 735 C<isXDIGIT_uvchr>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and 736 C<isXDIGIT_LC_utf8>. 737 738 =for apidoc Am|bool|isIDFIRST|char ch 739 Returns a boolean indicating whether the specified character can be the first 740 character of an identifier. This is very close to, but not quite the same as 741 the official Unicode property C<XID_Start>. The difference is that this 742 returns true only if the input character also matches L</isWORDCHAR>. 743 See the L<top of this section|/Character classification> for an explanation of 744 variants 745 C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uvchr>, C<isIDFIRST_utf8>, 746 C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>. 747 748 =for apidoc Am|bool|isIDCONT|char ch 749 Returns a boolean indicating whether the specified character can be the 750 second or succeeding character of an identifier. This is very close to, but 751 not quite the same as the official Unicode property C<XID_Continue>. The 752 difference is that this returns true only if the input character also matches 753 L</isWORDCHAR>. See the L<top of this section|/Character classification> for 754 an 755 explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uvchr>, 756 C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and 757 C<isIDCONT_LC_utf8>. 758 759 =head1 Miscellaneous Functions 760 761 =for apidoc Am|U8|READ_XDIGIT|char str* 762 Returns the value of an ASCII-range hex digit and advances the string pointer. 763 Behaviour is only well defined when isXDIGIT(*str) is true. 764 765 =head1 Character case changing 766 767 =for apidoc Am|U8|toUPPER|U8 ch 768 Converts the specified character to uppercase. If the input is anything but an 769 ASCII lowercase character, that input character itself is returned. Variant 770 C<toUPPER_A> is equivalent. 771 772 =for apidoc Am|UV|toUPPER_uvchr|UV cp|U8* s|STRLEN* lenp 773 Converts the code point C<cp> to its uppercase version, and 774 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code 775 point is interpreted as native if less than 256; otherwise as Unicode. Note 776 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 777 bytes since the uppercase version may be longer than the original character. 778 779 The first code point of the uppercased version is returned 780 (but note, as explained just above, that there may be more.) 781 782 =for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp 783 Converts the UTF-8 encoded character at C<p> to its uppercase version, and 784 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 785 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 786 bytes since the uppercase version may be longer than the original character. 787 788 The first code point of the uppercased version is returned 789 (but note, as explained just above, that there may be more.) 790 791 The input character at C<p> is assumed to be well-formed. 792 793 =for apidoc Am|U8|toFOLD|U8 ch 794 Converts the specified character to foldcase. If the input is anything but an 795 ASCII uppercase character, that input character itself is returned. Variant 796 C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full 797 Latin1 range, as the full generality of L</toFOLD_uvchr> is needed there.) 798 799 =for apidoc Am|UV|toFOLD_uvchr|UV cp|U8* s|STRLEN* lenp 800 Converts the code point C<cp> to its foldcase version, and 801 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code 802 point is interpreted as native if less than 256; otherwise as Unicode. Note 803 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 804 bytes since the foldcase version may be longer than the original character. 805 806 The first code point of the foldcased version is returned 807 (but note, as explained just above, that there may be more.) 808 809 =for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp 810 Converts the UTF-8 encoded character at C<p> to its foldcase version, and 811 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 812 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 813 bytes since the foldcase version may be longer than the original character. 814 815 The first code point of the foldcased version is returned 816 (but note, as explained just above, that there may be more.) 817 818 The input character at C<p> is assumed to be well-formed. 819 820 =for apidoc Am|U8|toLOWER|U8 ch 821 Converts the specified character to lowercase. If the input is anything but an 822 ASCII uppercase character, that input character itself is returned. Variant 823 C<toLOWER_A> is equivalent. 824 825 =for apidoc Am|U8|toLOWER_L1|U8 ch 826 Converts the specified Latin1 character to lowercase. The results are 827 undefined if the input doesn't fit in a byte. 828 829 =for apidoc Am|U8|toLOWER_LC|U8 ch 830 Converts the specified character to lowercase using the current locale's rules, 831 if possible; otherwise returns the input character itself. 832 833 =for apidoc Am|UV|toLOWER_uvchr|UV cp|U8* s|STRLEN* lenp 834 Converts the code point C<cp> to its lowercase version, and 835 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code 836 point is interpreted as native if less than 256; otherwise as Unicode. Note 837 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 838 bytes since the lowercase version may be longer than the original character. 839 840 The first code point of the lowercased version is returned 841 (but note, as explained just above, that there may be more.) 842 843 =for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp 844 Converts the UTF-8 encoded character at C<p> to its lowercase version, and 845 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 846 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 847 bytes since the lowercase version may be longer than the original character. 848 849 The first code point of the lowercased version is returned 850 (but note, as explained just above, that there may be more.) 851 852 The input character at C<p> is assumed to be well-formed. 853 854 =for apidoc Am|U8|toTITLE|U8 ch 855 Converts the specified character to titlecase. If the input is anything but an 856 ASCII lowercase character, that input character itself is returned. Variant 857 C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 858 range, as the full generality of L</toTITLE_uvchr> is needed there. Titlecase is 859 not a concept used in locale handling, so there is no functionality for that.) 860 861 =for apidoc Am|UV|toTITLE_uvchr|UV cp|U8* s|STRLEN* lenp 862 Converts the code point C<cp> to its titlecase version, and 863 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code 864 point is interpreted as native if less than 256; otherwise as Unicode. Note 865 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 866 bytes since the titlecase version may be longer than the original character. 867 868 The first code point of the titlecased version is returned 869 (but note, as explained just above, that there may be more.) 870 871 =for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp 872 Converts the UTF-8 encoded character at C<p> to its titlecase version, and 873 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 874 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 875 bytes since the titlecase version may be longer than the original character. 876 877 The first code point of the titlecased version is returned 878 (but note, as explained just above, that there may be more.) 879 880 The input character at C<p> is assumed to be well-formed. 881 882 =cut 883 884 XXX Still undocumented isVERTWS_uvchr and _utf8; it's unclear what their names 885 really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change. 886 887 Note that these macros are repeated in Devel::PPPort, so should also be 888 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc 889 890 */ 891 892 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64 893 * is known only in the perl core, and this macro can be called from outside 894 * that */ 895 #ifdef HAS_QUAD 896 # define WIDEST_UTYPE U64TYPE 897 #else 898 # define WIDEST_UTYPE U32 899 #endif 900 901 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in 902 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no 903 * bits of information are lost even on a 64-bit machine, but to get the 904 * compiler to optimize it out if possible. This is because Configure makes 905 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the 906 * sizeof() guarantees that this evaluates to a constant true at compile time. 907 * 908 * For Coverity, be always true, because otherwise Coverity thinks 909 * it finds several expressions that are always true, independent 910 * of operands. Well, they are, but that is kind of the point. 911 */ 912 #ifndef __COVERITY__ 913 #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF)) 914 #else 915 #define FITS_IN_8_BITS(c) (1) 916 #endif 917 918 #ifdef EBCDIC 919 # ifndef _ALL_SOURCE 920 /* The native libc isascii() et.al. functions return the wrong results 921 * on at least z/OS unless this is defined. */ 922 # error _ALL_SOURCE should probably be defined 923 # endif 924 #else 925 /* There is a simple definition of ASCII for ASCII platforms. But the 926 * EBCDIC one isn't so simple, so is defined using table look-up like the 927 * other macros below. 928 * The '| 0' part ensures that c is an integer (and not e.g. a pointer) */ 929 # define isASCII(c) ((WIDEST_UTYPE)((c) | 0) < 128) 930 #endif 931 932 /* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0, 933 * and the 8 possible permutations of those bits exactly comprise the 8 octal 934 * digits */ 935 #define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0') 936 937 #ifdef H_PERL /* If have access to perl.h, lookup in its table */ 938 939 /* Character class numbers. For internal core Perl use only. The ones less 940 * than 32 are used in PL_charclass[] and the ones up through the one that 941 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and 942 * related files. PL_charclass ones use names used in l1_char_class_tab.h but 943 * their actual definitions are here. If that file has a name not used here, 944 * it won't compile. 945 * 946 * The first group of these is ordered in what I (khw) estimate to be the 947 * frequency of their use. This gives a slight edge to exiting a loop earlier 948 * (in reginclass() in regexec.c) */ 949 # define _CC_WORDCHAR 0 /* \w and [:word:] */ 950 # define _CC_DIGIT 1 /* \d and [:digit:] */ 951 # define _CC_ALPHA 2 /* [:alpha:] */ 952 # define _CC_LOWER 3 /* [:lower:] */ 953 # define _CC_UPPER 4 /* [:upper:] */ 954 # define _CC_PUNCT 5 /* [:punct:] */ 955 # define _CC_PRINT 6 /* [:print:] */ 956 # define _CC_ALPHANUMERIC 7 /* [:alnum:] */ 957 # define _CC_GRAPH 8 /* [:graph:] */ 958 # define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */ 959 960 #define _FIRST_NON_SWASH_CC 10 961 /* The character classes above are implemented with swashes. The second group 962 * (just below) contains the ones implemented without. These are also sorted 963 * in rough order of the frequency of their use, except that \v should be last, 964 * as it isn't a real Posix character class, and some (small) inefficiencies in 965 * regular expression handling would be introduced by putting it in the middle 966 * of those that are. Also, cntrl and ascii come after the others as it may be 967 * useful to group these which have no members that match above Latin1, (or 968 * above ASCII in the latter case) */ 969 970 # define _CC_SPACE 10 /* \s, [:space:] */ 971 # define _CC_BLANK 11 /* [:blank:] */ 972 # define _CC_XDIGIT 12 /* [:xdigit:] */ 973 # define _CC_CNTRL 13 /* [:cntrl:] */ 974 # define _CC_ASCII 14 /* [:ascii:] */ 975 # define _CC_VERTSPACE 15 /* \v */ 976 977 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE 978 979 /* The members of the third group below do not need to be coordinated with data 980 * structures in regcomp.[ch] and regexec.c. */ 981 # define _CC_IDFIRST 16 982 # define _CC_CHARNAME_CONT 17 983 # define _CC_NONLATIN1_FOLD 18 984 # define _CC_NONLATIN1_SIMPLE_FOLD 19 985 # define _CC_QUOTEMETA 20 986 # define _CC_NON_FINAL_FOLD 21 987 # define _CC_IS_IN_SOME_FOLD 22 988 # define _CC_MNEMONIC_CNTRL 23 989 990 /* This next group is only used on EBCDIC platforms, so theoretically could be 991 * shared with something entirely different that's only on ASCII platforms */ 992 # define _CC_UTF8_START_BYTE_IS_FOR_AT_LEAST_SURROGATE 28 993 # define _CC_UTF8_IS_START 29 994 # define _CC_UTF8_IS_DOWNGRADEABLE_START 30 995 # define _CC_UTF8_IS_CONTINUATION 31 996 /* Unused: 24-27 997 * If more bits are needed, one could add a second word for non-64bit 998 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd 999 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it 1000 * is used only for optimization (as of this writing), and differs in the 1001 * Latin1 range from the ALPHA bit only in two relatively unimportant 1002 * characters: the masculine and feminine ordinal indicators, so removing it 1003 * would just cause /i regexes which match them to run less efficiently. 1004 * Similarly the EBCDIC-only bits are used just for speed, and could be 1005 * replaced by other means */ 1006 1007 #if defined(PERL_CORE) || defined(PERL_EXT) 1008 /* An enum version of the character class numbers, to help compilers 1009 * optimize */ 1010 typedef enum { 1011 _CC_ENUM_ALPHA = _CC_ALPHA, 1012 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC, 1013 _CC_ENUM_ASCII = _CC_ASCII, 1014 _CC_ENUM_BLANK = _CC_BLANK, 1015 _CC_ENUM_CASED = _CC_CASED, 1016 _CC_ENUM_CNTRL = _CC_CNTRL, 1017 _CC_ENUM_DIGIT = _CC_DIGIT, 1018 _CC_ENUM_GRAPH = _CC_GRAPH, 1019 _CC_ENUM_LOWER = _CC_LOWER, 1020 _CC_ENUM_PRINT = _CC_PRINT, 1021 _CC_ENUM_PUNCT = _CC_PUNCT, 1022 _CC_ENUM_SPACE = _CC_SPACE, 1023 _CC_ENUM_UPPER = _CC_UPPER, 1024 _CC_ENUM_VERTSPACE = _CC_VERTSPACE, 1025 _CC_ENUM_WORDCHAR = _CC_WORDCHAR, 1026 _CC_ENUM_XDIGIT = _CC_XDIGIT 1027 } _char_class_number; 1028 #endif 1029 1030 #define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC 1031 #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1) 1032 1033 #if defined(PERL_IN_UTF8_C) \ 1034 || defined(PERL_IN_REGCOMP_C) \ 1035 || defined(PERL_IN_REGEXEC_C) 1036 # if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \ 1037 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \ 1038 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9 1039 #error Need to adjust order of swash_property_names[] 1040 # endif 1041 1042 /* This is declared static in each of the few files that this is #defined for 1043 * to keep them from being publicly accessible. Hence there is a small amount 1044 * of wasted space */ 1045 1046 static const char* const swash_property_names[] = { 1047 "XPosixWord", 1048 "XPosixDigit", 1049 "XPosixAlpha", 1050 "XPosixLower", 1051 "XPosixUpper", 1052 "XPosixPunct", 1053 "XPosixPrint", 1054 "XPosixAlnum", 1055 "XPosixGraph", 1056 "Cased" 1057 }; 1058 #endif 1059 1060 START_EXTERN_C 1061 # ifdef DOINIT 1062 EXTCONST U32 PL_charclass[] = { 1063 # include "l1_char_class_tab.h" 1064 }; 1065 1066 # else /* ! DOINIT */ 1067 EXTCONST U32 PL_charclass[]; 1068 # endif 1069 END_EXTERN_C 1070 1071 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */ 1072 # define _CC_mask(classnum) (1U << (classnum)) 1073 1074 /* For internal core Perl use only: the base macro for defining macros like 1075 * isALPHA */ 1076 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \ 1077 && (PL_charclass[(U8) (c)] & _CC_mask(classnum))) 1078 1079 /* The mask for the _A versions of the macros; it just adds in the bit for 1080 * ASCII. */ 1081 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII)) 1082 1083 /* For internal core Perl use only: the base macro for defining macros like 1084 * isALPHA_A. The foo_A version makes sure that both the desired bit and 1085 * the ASCII bit are present */ 1086 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \ 1087 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \ 1088 == _CC_mask_A(classnum))) 1089 1090 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA) 1091 # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC) 1092 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK) 1093 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL) 1094 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT) /* No non-ASCII digits */ 1095 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH) 1096 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER) 1097 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT) 1098 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT) 1099 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE) 1100 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER) 1101 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR) 1102 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) /* No non-ASCII xdigits 1103 */ 1104 # define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST) 1105 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA) 1106 # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC) 1107 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK) 1108 1109 /* continuation character for legal NAME in \N{NAME} */ 1110 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT) 1111 1112 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL) 1113 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH) 1114 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER) 1115 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT) 1116 # define isPSXSPC_L1(c) isSPACE_L1(c) 1117 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT) 1118 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE) 1119 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER) 1120 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR) 1121 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST) 1122 1123 # ifdef EBCDIC 1124 # define isASCII(c) _generic_isCC(c, _CC_ASCII) 1125 # endif 1126 1127 /* Participates in a single-character fold with a character above 255 */ 1128 # define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_SIMPLE_FOLD))) 1129 1130 /* Like the above, but also can be part of a multi-char fold */ 1131 # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_FOLD))) 1132 1133 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA) 1134 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ 1135 _generic_isCC(c, _CC_NON_FINAL_FOLD) 1136 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ 1137 _generic_isCC(c, _CC_IS_IN_SOME_FOLD) 1138 # define _IS_MNEMONIC_CNTRL_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ 1139 _generic_isCC(c, _CC_MNEMONIC_CNTRL) 1140 #else /* else we don't have perl.h H_PERL */ 1141 1142 /* If we don't have perl.h, we are compiling a utility program. Below we 1143 * hard-code various macro definitions that wouldn't otherwise be available 1144 * to it. Most are coded based on first principals. First some ones common 1145 * to both ASCII and EBCDIC */ 1146 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0') 1147 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t') 1148 # define isSPACE_A(c) (isBLANK_A(c) \ 1149 || (c) == '\n' \ 1150 || (c) == '\r' \ 1151 || (c) == '\v' \ 1152 || (c) == '\f') 1153 # ifdef EBCDIC /* There are gaps between 'i' and 'j'; 'r' and 's'. Same 1154 for uppercase. This is ordered to exclude most things 1155 early */ 1156 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z' \ 1157 && ((c) <= 'i' \ 1158 || ((c) >= 'j' && (c) <= 'r') \ 1159 || (c) >= 's')) 1160 # define isUPPER_A(c) ((c) >= 'A' && (c) <= 'Z' \ 1161 && ((c) <= 'I' \ 1162 || ((c) >= 'J' && (c) <= 'R') \ 1163 || (c) >= 'S')) 1164 # else /* ASCII platform. */ 1165 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z') 1166 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A') 1167 # endif 1168 1169 /* Some more ASCII, non-ASCII common definitions */ 1170 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c)) 1171 # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c)) 1172 # define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_') 1173 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_') 1174 # define isXDIGIT_A(c) (isDIGIT_A(c) \ 1175 || ((c) >= 'a' && (c) <= 'f') \ 1176 || ((c) <= 'F' && (c) >= 'A')) 1177 1178 # ifdef EBCDIC 1179 # define isPUNCT_A(c) ((c) == '-' || (c) == '!' || (c) == '"' \ 1180 || (c) == '#' || (c) == '$' || (c) == '%' \ 1181 || (c) == '&' || (c) == '\'' || (c) == '(' \ 1182 || (c) == ')' || (c) == '*' || (c) == '+' \ 1183 || (c) == ',' || (c) == '.' || (c) == '/' \ 1184 || (c) == ':' || (c) == ';' || (c) == '<' \ 1185 || (c) == '=' || (c) == '>' || (c) == '?' \ 1186 || (c) == '@' || (c) == '[' || (c) == '\\' \ 1187 || (c) == ']' || (c) == '^' || (c) == '_' \ 1188 || (c) == '`' || (c) == '{' || (c) == '|' \ 1189 || (c) == '}' || (c) == '~') 1190 # define isGRAPH_A(c) (isALPHANUMERIC_A(c) || isPUNCT_A(c)) 1191 # define isPRINT_A(c) (isGRAPH_A(c) || (c) == ' ') 1192 1193 # ifdef QUESTION_MARK_CTRL 1194 # define _isQMC(c) ((c) == QUESTION_MARK_CTRL) 1195 # else 1196 # define _isQMC(c) 0 1197 # endif 1198 1199 /* I (khw) can't think of a way to define all the ASCII controls 1200 * without resorting to a libc (locale-sensitive) call. But we know 1201 * that all controls but the question-mark one are in the range 0-0x3f. 1202 * This makes sure that all the controls that have names are included, 1203 * and all controls that are also considered ASCII in the locale. This 1204 * may include more or fewer than what it actually should, but the 1205 * wrong ones are less-important controls, so likely won't impact 1206 * things (keep in mind that this is compiled only if perl.h isn't 1207 * available). The question mark control is included if available */ 1208 # define isCNTRL_A(c) (((c) < 0x40 && isascii(c)) \ 1209 || (c) == '\0' || (c) == '\a' || (c) == '\b' \ 1210 || (c) == '\f' || (c) == '\n' || (c) == '\r' \ 1211 || (c) == '\t' || (c) == '\v' || _isQMC(c)) 1212 1213 # define isASCII(c) (isCNTRL_A(c) || isPRINT_A(c)) 1214 # else /* ASCII platform; things are simpler, and isASCII has already 1215 been defined */ 1216 # define isGRAPH_A(c) (((c) > ' ' && (c) < 127)) 1217 # define isPRINT_A(c) (isGRAPH_A(c) || (c) == ' ') 1218 # define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c))) 1219 # define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c))) 1220 # endif 1221 1222 /* The _L1 macros may be unnecessary for the utilities; I (khw) added them 1223 * during debugging, and it seems best to keep them. We may be called 1224 * without NATIVE_TO_LATIN1 being defined. On ASCII platforms, it doesn't 1225 * do anything anyway, so make it not a problem */ 1226 # if ! defined(EBCDIC) && ! defined(NATIVE_TO_LATIN1) 1227 # define NATIVE_TO_LATIN1(ch) (ch) 1228 # endif 1229 # define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c)) 1230 # define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c)) 1231 # define isBLANK_L1(c) (isBLANK_A(c) \ 1232 || (FITS_IN_8_BITS(c) \ 1233 && NATIVE_TO_LATIN1((U8) c) == 0xA0)) 1234 # define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c))) 1235 # define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c))) 1236 # define isLOWER_L1(c) (isLOWER_A(c) \ 1237 || (FITS_IN_8_BITS(c) \ 1238 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \ 1239 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \ 1240 || NATIVE_TO_LATIN1((U8) c) == 0xAA \ 1241 || NATIVE_TO_LATIN1((U8) c) == 0xBA \ 1242 || NATIVE_TO_LATIN1((U8) c) == 0xB5))) 1243 # define isPRINT_L1(c) (isPRINT_A(c) \ 1244 || (FITS_IN_8_BITS(c) \ 1245 && NATIVE_TO_LATIN1((U8) c) >= 0xA0)) 1246 # define isPUNCT_L1(c) (isPUNCT_A(c) \ 1247 || (FITS_IN_8_BITS(c) \ 1248 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \ 1249 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \ 1250 || NATIVE_TO_LATIN1((U8) c) == 0xAB \ 1251 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \ 1252 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \ 1253 || NATIVE_TO_LATIN1((U8) c) == 0xBB \ 1254 || NATIVE_TO_LATIN1((U8) c) == 0xBF))) 1255 # define isSPACE_L1(c) (isSPACE_A(c) \ 1256 || (FITS_IN_8_BITS(c) \ 1257 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \ 1258 || NATIVE_TO_LATIN1((U8) c) == 0xA0))) 1259 # define isUPPER_L1(c) (isUPPER_A(c) \ 1260 || (FITS_IN_8_BITS(c) \ 1261 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \ 1262 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \ 1263 && NATIVE_TO_LATIN1((U8) c) != 0xD7))) 1264 # define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c)) 1265 # define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_') 1266 # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \ 1267 || isBLANK_L1(c) \ 1268 || (c) == '-' \ 1269 || (c) == '(' \ 1270 || (c) == ')') 1271 /* The following are not fully accurate in the above-ASCII range. I (khw) 1272 * don't think it's necessary to be so for the purposes where this gets 1273 * compiled */ 1274 # define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c)) 1275 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c) 1276 1277 /* And these aren't accurate at all. They are useful only for above 1278 * Latin1, which utilities and bootstrapping don't deal with */ 1279 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0 1280 # define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0 1281 # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0 1282 1283 /* Many of the macros later in this file are defined in terms of these. By 1284 * implementing them with a function, which converts the class number into 1285 * a call to the desired macro, all of the later ones work. However, that 1286 * function won't be actually defined when building a utility program (no 1287 * perl.h), and so a compiler error will be generated if one is attempted 1288 * to be used. And the above-Latin1 code points require Unicode tables to 1289 * be present, something unlikely to be the case when bootstrapping */ 1290 # define _generic_isCC(c, classnum) \ 1291 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE)) 1292 # define _generic_isCC_A(c, classnum) \ 1293 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE)) 1294 #endif /* End of no perl.h H_PERL */ 1295 1296 #define isALPHANUMERIC(c) isALPHANUMERIC_A(c) 1297 #define isALPHA(c) isALPHA_A(c) 1298 #define isASCII_A(c) isASCII(c) 1299 #define isASCII_L1(c) isASCII(c) 1300 #define isBLANK(c) isBLANK_A(c) 1301 #define isCNTRL(c) isCNTRL_A(c) 1302 #define isDIGIT(c) isDIGIT_A(c) 1303 #define isGRAPH(c) isGRAPH_A(c) 1304 #define isIDFIRST(c) isIDFIRST_A(c) 1305 #define isLOWER(c) isLOWER_A(c) 1306 #define isPRINT(c) isPRINT_A(c) 1307 #define isPSXSPC_A(c) isSPACE_A(c) 1308 #define isPSXSPC(c) isPSXSPC_A(c) 1309 #define isPSXSPC_L1(c) isSPACE_L1(c) 1310 #define isPUNCT(c) isPUNCT_A(c) 1311 #define isSPACE(c) isSPACE_A(c) 1312 #define isUPPER(c) isUPPER_A(c) 1313 #define isWORDCHAR(c) isWORDCHAR_A(c) 1314 #define isXDIGIT(c) isXDIGIT_A(c) 1315 1316 /* ASCII casing. These could also be written as 1317 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c)) 1318 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c)) 1319 which uses table lookup and mask instead of subtraction. (This would 1320 work because the _MOD does not apply in the ASCII range) */ 1321 #define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c)) 1322 #define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c)) 1323 1324 /* In the ASCII range, these are equivalent to what they're here defined to be. 1325 * But by creating these definitions, other code doesn't have to be aware of 1326 * this detail */ 1327 #define toFOLD(c) toLOWER(c) 1328 #define toTITLE(c) toUPPER(c) 1329 1330 #define toLOWER_A(c) toLOWER(c) 1331 #define toUPPER_A(c) toUPPER(c) 1332 #define toFOLD_A(c) toFOLD(c) 1333 #define toTITLE_A(c) toTITLE(c) 1334 1335 /* Use table lookup for speed; returns the input itself if is out-of-range */ 1336 #define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \ 1337 ? (c) \ 1338 : PL_latin1_lc[ (U8) (c) ]) 1339 #define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */ 1340 1341 /* Modified uc. Is correct uc except for three non-ascii chars which are 1342 * all mapped to one of them, and these need special handling; returns the 1343 * input itself if is out-of-range */ 1344 #define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \ 1345 ? (c) \ 1346 : PL_mod_latin1_uc[ (U8) (c) ]) 1347 #define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale 1348 1349 /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */ 1350 1351 /* For internal core Perl use only: the base macro for defining macros like 1352 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point 1353 * (0-255) to check. In a UTF-8 locale, the result is the same as calling 1354 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like 1355 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8 1356 * locales, the code to actually do the test this is passed in 'non_utf8'. If 1357 * 'c' is above 255, 0 is returned. For accessing the full range of possible 1358 * code points under locale rules, use the macros based on _generic_LC_uvchr 1359 * instead of this. */ 1360 #define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \ 1361 (! FITS_IN_8_BITS(c) \ 1362 ? 0 \ 1363 : IN_UTF8_CTYPE_LOCALE \ 1364 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \ 1365 : cBOOL(non_utf8)) 1366 1367 /* For internal core Perl use only: a helper macro for defining macros like 1368 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to 1369 * actually do this test is passed in 'non_utf8_func', which is called on 'c', 1370 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See 1371 * _generic_LC_base for more info */ 1372 #define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \ 1373 _generic_LC_base(c,utf8_locale_classnum, \ 1374 non_utf8_func( (_LC_CAST) (c))) 1375 1376 /* For internal core Perl use only: like _generic_LC, but also returns TRUE if 1377 * 'c' is the platform's native underscore character */ 1378 #define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \ 1379 _generic_LC_base(c, utf8_locale_classnum, \ 1380 (non_utf8_func( (_LC_CAST) (c)) \ 1381 || (char)(c) == '_')) 1382 1383 /* These next three are also for internal core Perl use only: case-change 1384 * helper macros */ 1385 #define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \ 1386 ? (c) \ 1387 : (IN_UTF8_CTYPE_LOCALE) \ 1388 ? PL_latin1_lc[ (U8) (c) ] \ 1389 : (cast)function((cast)(c))) 1390 1391 /* Note that the result can be larger than a byte in a UTF-8 locale. It 1392 * returns a single value, so can't adequately return the upper case of LATIN 1393 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two 1394 * values "SS"); instead it asserts against that under DEBUGGING, and 1395 * otherwise returns its input */ 1396 #define _generic_toUPPER_LC(c, function, cast) \ 1397 (! FITS_IN_8_BITS(c) \ 1398 ? (c) \ 1399 : ((! IN_UTF8_CTYPE_LOCALE) \ 1400 ? (cast)function((cast)(c)) \ 1401 : ((((U8)(c)) == MICRO_SIGN) \ 1402 ? GREEK_CAPITAL_LETTER_MU \ 1403 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \ 1404 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \ 1405 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \ 1406 ? (__ASSERT_(0) (c)) \ 1407 : PL_mod_latin1_uc[ (U8) (c) ]))))) 1408 1409 /* Note that the result can be larger than a byte in a UTF-8 locale. It 1410 * returns a single value, so can't adequately return the fold case of LATIN 1411 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two 1412 * values "ss"); instead it asserts against that under DEBUGGING, and 1413 * otherwise returns its input */ 1414 #define _generic_toFOLD_LC(c, function, cast) \ 1415 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \ 1416 ? GREEK_SMALL_LETTER_MU \ 1417 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \ 1418 || (c) != LATIN_SMALL_LETTER_SHARP_S) \ 1419 _generic_toLOWER_LC(c, function, cast))) 1420 1421 /* Use the libc versions for these if available. */ 1422 #if defined(HAS_ISASCII) 1423 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c))) 1424 #else 1425 # define isASCII_LC(c) isASCII(c) 1426 #endif 1427 1428 #if defined(HAS_ISBLANK) 1429 # define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank) 1430 #else /* Unlike isASCII, varies if in a UTF-8 locale */ 1431 # define isBLANK_LC(c) ((IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)) 1432 #endif 1433 1434 #define _LC_CAST U8 1435 1436 #ifdef WIN32 1437 /* The Windows functions don't bother to follow the POSIX standard, which 1438 * for example says that something can't both be a printable and a control. 1439 * But Windows treats the \t control as a printable, and does such things 1440 * as making superscripts into both digits and punctuation. This tames 1441 * these flaws by assuming that the definitions of both controls and space 1442 * are correct, and then making sure that other definitions don't have 1443 * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc. 1444 * Not all possible weirdnesses are checked for, just the ones that were 1445 * detected on actual Microsoft code pages */ 1446 1447 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl) 1448 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace) 1449 1450 # define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) \ 1451 && isALPHANUMERIC_LC(c)) 1452 # define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && \ 1453 ! isPUNCT_LC(c)) 1454 # define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && \ 1455 isALPHANUMERIC_LC(c)) 1456 # define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c)) 1457 # define isIDFIRST_LC(c) (((c) == '_') \ 1458 || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c))) 1459 # define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c)) 1460 # define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c)) 1461 # define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c)) 1462 # define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c)) 1463 # define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c)) 1464 # define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) \ 1465 && isALPHANUMERIC_LC(c)) 1466 1467 # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8) 1468 # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8) 1469 # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8) 1470 1471 #elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII)) 1472 /* For most other platforms */ 1473 1474 # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha) 1475 # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum) 1476 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl) 1477 # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit) 1478 # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph) 1479 # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha) 1480 # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower) 1481 # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint) 1482 # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct) 1483 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace) 1484 # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper) 1485 # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum) 1486 # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit) 1487 1488 1489 # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8) 1490 # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8) 1491 # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8) 1492 1493 #else /* The final fallback position */ 1494 1495 # define isALPHA_LC(c) (isascii(c) && isalpha(c)) 1496 # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c)) 1497 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c)) 1498 # define isDIGIT_LC(c) (isascii(c) && isdigit(c)) 1499 # define isGRAPH_LC(c) (isascii(c) && isgraph(c)) 1500 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_')) 1501 # define isLOWER_LC(c) (isascii(c) && islower(c)) 1502 # define isPRINT_LC(c) (isascii(c) && isprint(c)) 1503 # define isPUNCT_LC(c) (isascii(c) && ispunct(c)) 1504 # define isSPACE_LC(c) (isascii(c) && isspace(c)) 1505 # define isUPPER_LC(c) (isascii(c) && isupper(c)) 1506 # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_')) 1507 # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c)) 1508 1509 # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c)) 1510 # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c)) 1511 # define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c)) 1512 1513 #endif 1514 1515 #define isIDCONT(c) isWORDCHAR(c) 1516 #define isIDCONT_A(c) isWORDCHAR_A(c) 1517 #define isIDCONT_L1(c) isWORDCHAR_L1(c) 1518 #define isIDCONT_LC(c) isWORDCHAR_LC(c) 1519 #define isPSXSPC_LC(c) isSPACE_LC(c) 1520 1521 /* For internal core Perl use only: the base macros for defining macros like 1522 * isALPHA_uvchr. 'c' is the code point to check. 'classnum' is the POSIX class 1523 * number defined earlier in this file. _generic_uvchr() is used for POSIX 1524 * classes where there is a macro or function 'above_latin1' that takes the 1525 * single argument 'c' and returns the desired value. These exist for those 1526 * classes which have simple definitions, avoiding the overhead of a hash 1527 * lookup or inversion list binary search. _generic_swash_uvchr() can be used 1528 * for classes where that overhead is faster than a direct lookup. 1529 * _generic_uvchr() won't compile if 'c' isn't unsigned, as it won't match the 1530 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so 1531 * have duplicate checks here, so could create versions of the macros that 1532 * don't, but experiments show that gcc optimizes them out anyway. */ 1533 1534 /* Note that all ignore 'use bytes' */ 1535 #define _generic_uvchr(classnum, above_latin1, c) ((c) < 256 \ 1536 ? _generic_isCC(c, classnum) \ 1537 : above_latin1(c)) 1538 #define _generic_swash_uvchr(classnum, c) ((c) < 256 \ 1539 ? _generic_isCC(c, classnum) \ 1540 : _is_uni_FOO(classnum, c)) 1541 #define isALPHA_uvchr(c) _generic_swash_uvchr(_CC_ALPHA, c) 1542 #define isALPHANUMERIC_uvchr(c) _generic_swash_uvchr(_CC_ALPHANUMERIC, c) 1543 #define isASCII_uvchr(c) isASCII(c) 1544 #define isBLANK_uvchr(c) _generic_uvchr(_CC_BLANK, is_HORIZWS_cp_high, c) 1545 #define isCNTRL_uvchr(c) isCNTRL_L1(c) /* All controls are in Latin1 */ 1546 #define isDIGIT_uvchr(c) _generic_swash_uvchr(_CC_DIGIT, c) 1547 #define isGRAPH_uvchr(c) _generic_swash_uvchr(_CC_GRAPH, c) 1548 #define isIDCONT_uvchr(c) _generic_uvchr(_CC_WORDCHAR, _is_uni_perl_idcont, c) 1549 #define isIDFIRST_uvchr(c) _generic_uvchr(_CC_IDFIRST, _is_uni_perl_idstart, c) 1550 #define isLOWER_uvchr(c) _generic_swash_uvchr(_CC_LOWER, c) 1551 #define isPRINT_uvchr(c) _generic_swash_uvchr(_CC_PRINT, c) 1552 1553 #define isPUNCT_uvchr(c) _generic_swash_uvchr(_CC_PUNCT, c) 1554 #define isSPACE_uvchr(c) _generic_uvchr(_CC_SPACE, is_XPERLSPACE_cp_high, c) 1555 #define isPSXSPC_uvchr(c) isSPACE_uvchr(c) 1556 1557 #define isUPPER_uvchr(c) _generic_swash_uvchr(_CC_UPPER, c) 1558 #define isVERTWS_uvchr(c) _generic_uvchr(_CC_VERTSPACE, is_VERTWS_cp_high, c) 1559 #define isWORDCHAR_uvchr(c) _generic_swash_uvchr(_CC_WORDCHAR, c) 1560 #define isXDIGIT_uvchr(c) _generic_uvchr(_CC_XDIGIT, is_XDIGIT_cp_high, c) 1561 1562 #define toFOLD_uvchr(c,s,l) to_uni_fold(c,s,l) 1563 #define toLOWER_uvchr(c,s,l) to_uni_lower(c,s,l) 1564 #define toTITLE_uvchr(c,s,l) to_uni_title(c,s,l) 1565 #define toUPPER_uvchr(c,s,l) to_uni_upper(c,s,l) 1566 1567 /* For backwards compatibility, even though '_uni' should mean official Unicode 1568 * code points, in Perl it means native for those below 256 */ 1569 #define isALPHA_uni(c) isALPHA_uvchr(c) 1570 #define isALPHANUMERIC_uni(c) isALPHANUMERIC_uvchr(c) 1571 #define isASCII_uni(c) isASCII_uvchr(c) 1572 #define isBLANK_uni(c) isBLANK_uvchr(c) 1573 #define isCNTRL_uni(c) isCNTRL_uvchr(c) 1574 #define isDIGIT_uni(c) isDIGIT_uvchr(c) 1575 #define isGRAPH_uni(c) isGRAPH_uvchr(c) 1576 #define isIDCONT_uni(c) isIDCONT_uvchr(c) 1577 #define isIDFIRST_uni(c) isIDFIRST_uvchr(c) 1578 #define isLOWER_uni(c) isLOWER_uvchr(c) 1579 #define isPRINT_uni(c) isPRINT_uvchr(c) 1580 #define isPUNCT_uni(c) isPUNCT_uvchr(c) 1581 #define isSPACE_uni(c) isSPACE_uvchr(c) 1582 #define isPSXSPC_uni(c) isPSXSPC_uvchr(c) 1583 #define isUPPER_uni(c) isUPPER_uvchr(c) 1584 #define isVERTWS_uni(c) isVERTWS_uvchr(c) 1585 #define isWORDCHAR_uni(c) isWORDCHAR_uvchr(c) 1586 #define isXDIGIT_uni(c) isXDIGIT_uvchr(c) 1587 #define toFOLD_uni(c,s,l) toFOLD_uvchr(c,s,l) 1588 #define toLOWER_uni(c,s,l) toLOWER_uvchr(c,s,l) 1589 #define toTITLE_uni(c,s,l) toTITLE_uvchr(c,s,l) 1590 #define toUPPER_uni(c,s,l) toUPPER_uvchr(c,s,l) 1591 1592 /* For internal core Perl use only: the base macros for defining macros like 1593 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code 1594 * point, not just 0-255. Like _generic_uvchr, there are two versions, one for 1595 * simple class definitions; the other for more complex. These are like 1596 * _generic_uvchr, so see it for more info. */ 1597 #define _generic_LC_uvchr(latin1, above_latin1, c) \ 1598 (c < 256 ? latin1(c) : above_latin1(c)) 1599 #define _generic_LC_swash_uvchr(latin1, classnum, c) \ 1600 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c)) 1601 1602 #define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c) 1603 #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \ 1604 _CC_ALPHANUMERIC, c) 1605 #define isASCII_LC_uvchr(c) isASCII_LC(c) 1606 #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, \ 1607 is_HORIZWS_cp_high, c) 1608 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0) 1609 #define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c) 1610 #define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c) 1611 #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \ 1612 _is_uni_perl_idcont, c) 1613 #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \ 1614 _is_uni_perl_idstart, c) 1615 #define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c) 1616 #define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c) 1617 #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) 1618 #define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c) 1619 #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \ 1620 is_XPERLSPACE_cp_high, c) 1621 #define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c) 1622 #define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \ 1623 _CC_WORDCHAR, c) 1624 #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, \ 1625 is_XDIGIT_cp_high, c) 1626 1627 #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c)) 1628 1629 /* For internal core Perl use only: the base macros for defining macros like 1630 * isALPHA_utf8. These are like the earlier defined macros, but take an input 1631 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use 1632 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the 1633 * 'utf8' parameter. This relies on the fact that ASCII characters have the 1634 * same representation whether utf8 or not. Note that it assumes that the utf8 1635 * has been validated, and ignores 'use bytes' */ 1636 #define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \ 1637 ? _generic_isCC(*(p), classnum) \ 1638 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ 1639 ? _generic_isCC( \ 1640 EIGHT_BIT_UTF8_TO_NATIVE(*(p), \ 1641 *((p)+1 )), \ 1642 classnum) \ 1643 : utf8) 1644 /* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 1645 * 'above_latin1' can be a macro */ 1646 #define _generic_func_utf8(classnum, above_latin1, p) \ 1647 _generic_utf8(classnum, p, above_latin1(p)) 1648 /* Like the above, but passes classnum to _isFOO_utf8(), instead of having an 1649 * 'above_latin1' parameter */ 1650 #define _generic_swash_utf8(classnum, p) \ 1651 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p)) 1652 1653 /* Like the above, but should be used only when it is known that there are no 1654 * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the 1655 * class is TRUE for. Hence it can skip the tests for this range. 1656 * 'above_latin1' should include its arguments */ 1657 #define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \ 1658 (UTF8_IS_INVARIANT(*(p)) \ 1659 ? _generic_isCC(*(p), classnum) \ 1660 : (UTF8_IS_ABOVE_LATIN1(*(p))) \ 1661 ? above_latin1 \ 1662 : 0) 1663 1664 /* NOTE that some of these macros have very similar ones in regcharclass.h. 1665 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()' 1666 * there, differing in name only by an underscore from the one here 1667 * 'isSPACE_utf8(). The difference is that the ones here are probably more 1668 * efficient and smaller, using an O(1) array lookup for Latin1-range code 1669 * points; the regcharclass.h ones are implemented as a series of 1670 * "if-else-if-else ..." */ 1671 1672 #define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p) 1673 #define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p) 1674 #define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under 1675 utf8, the non-utf8 macro works 1676 */ 1677 #define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p) 1678 1679 #ifdef EBCDIC 1680 /* Because all controls are UTF-8 invariants in EBCDIC, we can use this 1681 * more efficient macro instead of the more general one */ 1682 # define isCNTRL_utf8(p) isCNTRL_L1(*(p)) 1683 #else 1684 # define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0) 1685 #endif 1686 1687 #define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \ 1688 _is_utf8_FOO(_CC_DIGIT, p)) 1689 #define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p) 1690 #define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \ 1691 _is_utf8_perl_idcont, p) 1692 1693 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that 1694 * IDFIRST is an alnum. See 1695 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you 1696 * ever wanted to know about. (In the ASCII range, there isn't a difference.) 1697 * This used to be not the XID version, but we decided to go with the more 1698 * modern Unicode definition */ 1699 #define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \ 1700 _is_utf8_perl_idstart, p) 1701 1702 #define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p) 1703 #define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p) 1704 #define isPSXSPC_utf8(p) isSPACE_utf8(p) 1705 #define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p) 1706 #define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p) 1707 #define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p) 1708 #define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p) 1709 #define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p) 1710 #define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \ 1711 is_XDIGIT_high(p)) 1712 1713 #define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l) 1714 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l) 1715 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l) 1716 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l) 1717 1718 /* For internal core Perl use only: the base macros for defining macros like 1719 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point 1720 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in 1721 * 'macro' parameter */ 1722 #define _generic_LC_utf8(macro, p, utf8) \ 1723 (UTF8_IS_INVARIANT(*(p)) \ 1724 ? macro(*(p)) \ 1725 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ 1726 ? macro(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1)))\ 1727 : utf8) 1728 1729 #define _generic_LC_swash_utf8(macro, classnum, p) \ 1730 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p)) 1731 #define _generic_LC_func_utf8(macro, above_latin1, p) \ 1732 _generic_LC_utf8(macro, p, above_latin1(p)) 1733 1734 #define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \ 1735 _CC_ALPHANUMERIC, p) 1736 #define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p) 1737 #define isASCII_LC_utf8(p) isASCII_LC(*p) 1738 #define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, \ 1739 is_HORIZWS_high, p) 1740 #define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0) 1741 #define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p) 1742 #define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p) 1743 #define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, \ 1744 _is_utf8_perl_idcont, p) 1745 #define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, \ 1746 _is_utf8_perl_idstart, p) 1747 #define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p) 1748 #define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p) 1749 #define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) 1750 #define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p) 1751 #define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, \ 1752 is_XPERLSPACE_high, p) 1753 #define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p) 1754 #define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \ 1755 _CC_WORDCHAR, p) 1756 #define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, \ 1757 is_XDIGIT_high, p) 1758 1759 /* Macros for backwards compatibility and for completeness when the ASCII and 1760 * Latin1 values are identical */ 1761 #define isALPHAU(c) isALPHA_L1(c) 1762 #define isDIGIT_L1(c) isDIGIT_A(c) 1763 #define isOCTAL(c) isOCTAL_A(c) 1764 #define isOCTAL_L1(c) isOCTAL_A(c) 1765 #define isXDIGIT_L1(c) isXDIGIT_A(c) 1766 #define isALNUM(c) isWORDCHAR(c) 1767 #define isALNUMU(c) isWORDCHAR_L1(c) 1768 #define isALNUM_LC(c) isWORDCHAR_LC(c) 1769 #define isALNUM_uni(c) isWORDCHAR_uni(c) 1770 #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c) 1771 #define isALNUM_utf8(p) isWORDCHAR_utf8(p) 1772 #define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p) 1773 #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */ 1774 #define isALNUMC_L1(c) isALPHANUMERIC_L1(c) 1775 #define isALNUMC(c) isALPHANUMERIC(c) 1776 #define isALNUMC_LC(c) isALPHANUMERIC_LC(c) 1777 #define isALNUMC_uni(c) isALPHANUMERIC_uni(c) 1778 #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c) 1779 #define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p) 1780 #define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p) 1781 1782 /* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, 1783 * except that they don't necessarily mean the same characters, e.g. CTRL-D is 1784 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC. 1785 * '?' is special-cased on EBCDIC to APC, which is the control there that is 1786 * the outlier from the block that contains the other controls, just like 1787 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0 1788 * block. If it weren't special cased, it would yield a non-control. 1789 * The conversion works both ways, so toCTRL('D') is 4, and toCTRL(4) is D, 1790 * etc. */ 1791 #ifndef EBCDIC 1792 # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) toUPPER(((U8)(c))) ^ 64) 1793 #else 1794 # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ 1795 ((isPRINT_A(c)) \ 1796 ? (UNLIKELY((c) == '?') \ 1797 ? QUESTION_MARK_CTRL \ 1798 : (NATIVE_TO_LATIN1(toUPPER((U8) (c))) ^ 64)) \ 1799 : (UNLIKELY((c) == QUESTION_MARK_CTRL) \ 1800 ? '?' \ 1801 : (LATIN1_TO_NATIVE(((U8) (c)) ^ 64))))) 1802 #endif 1803 1804 /* Line numbers are unsigned, 32 bits. */ 1805 typedef U32 line_t; 1806 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */ 1807 1808 /* Helpful alias for version prescan */ 1809 #define is_LAX_VERSION(a,b) \ 1810 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL)) 1811 1812 #define is_STRICT_VERSION(a,b) \ 1813 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL)) 1814 1815 #define BADVERSION(a,b,c) \ 1816 if (b) { \ 1817 *b = c; \ 1818 } \ 1819 return a; 1820 1821 /* Converts a character known to represent a hexadecimal digit (0-9, A-F, or 1822 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer, 1823 * which is advanced. The input is validated only by an assert() in DEBUGGING 1824 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and 1825 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */ 1826 #define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \ 1827 ? (c) \ 1828 : ((c) + 9)))) 1829 #define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \ 1830 ? (*(s)++) \ 1831 : (*(s)++ + 9)))) 1832 1833 /* Converts a character known to represent an octal digit (0-7) to its numeric 1834 * value. The input is validated only by an assert() in DEBUGGING builds. In 1835 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */ 1836 #define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c))) 1837 1838 /* Efficiently returns a boolean as to if two native characters are equivalent 1839 * case-insenstively. At least one of the characters must be one of [A-Za-z]; 1840 * the ALPHA in the name is to remind you of that. This is asserted() in 1841 * DEBUGGING builds. Because [A-Za-z] are invariant under UTF-8, this macro 1842 * works (on valid input) for both non- and UTF-8-encoded bytes. 1843 * 1844 * When one of the inputs is a compile-time constant and gets folded by the 1845 * compiler, this reduces to an AND and a TEST. On both EBCDIC and ASCII 1846 * machines, 'A' and 'a' differ by a single bit; the same with the upper and 1847 * lower case of all other ASCII-range alphabetics. On ASCII platforms, they 1848 * are 32 apart; on EBCDIC, they are 64. At compile time, this uses an 1849 * exclusive 'or' to find that bit and then inverts it to form a mask, with 1850 * just a single 0, in the bit position where the upper- and lowercase differ. 1851 * */ 1852 #define isALPHA_FOLD_EQ(c1, c2) \ 1853 (__ASSERT_(isALPHA_A(c1) || isALPHA_A(c2)) \ 1854 ((c1) & ~('A' ^ 'a')) == ((c2) & ~('A' ^ 'a'))) 1855 #define isALPHA_FOLD_NE(c1, c2) (! isALPHA_FOLD_EQ((c1), (c2))) 1856 1857 /* 1858 =head1 Memory Management 1859 1860 =for apidoc Am|void|Newx|void* ptr|int nitems|type 1861 The XSUB-writer's interface to the C C<malloc> function. 1862 1863 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1864 1865 In 5.9.3, Newx() and friends replace the older New() API, and drops 1866 the first parameter, I<x>, a debug aid which allowed callers to identify 1867 themselves. This aid has been superseded by a new build option, 1868 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still 1869 there for use in XS modules supporting older perls. 1870 1871 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast 1872 The XSUB-writer's interface to the C C<malloc> function, with 1873 cast. See also C<L</Newx>>. 1874 1875 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1876 1877 =for apidoc Am|void|Newxz|void* ptr|int nitems|type 1878 The XSUB-writer's interface to the C C<malloc> function. The allocated 1879 memory is zeroed with C<memzero>. See also C<L</Newx>>. 1880 1881 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1882 1883 =for apidoc Am|void|Renew|void* ptr|int nitems|type 1884 The XSUB-writer's interface to the C C<realloc> function. 1885 1886 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1887 1888 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast 1889 The XSUB-writer's interface to the C C<realloc> function, with 1890 cast. 1891 1892 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1893 1894 =for apidoc Am|void|Safefree|void* ptr 1895 The XSUB-writer's interface to the C C<free> function. 1896 1897 This should B<ONLY> be used on memory obtained using L<"Newx"> and friends. 1898 1899 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type 1900 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the 1901 source, C<dest> is the destination, C<nitems> is the number of items, and 1902 C<type> is the type. Can do overlapping moves. See also C<L</Copy>>. 1903 1904 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type 1905 Like C<Move> but returns C<dest>. Useful 1906 for encouraging compilers to tail-call 1907 optimise. 1908 1909 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type 1910 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the 1911 source, C<dest> is the destination, C<nitems> is the number of items, and 1912 C<type> is the type. May fail on overlapping copies. See also C<L</Move>>. 1913 1914 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type 1915 1916 Like C<Copy> but returns C<dest>. Useful 1917 for encouraging compilers to tail-call 1918 optimise. 1919 1920 =for apidoc Am|void|Zero|void* dest|int nitems|type 1921 1922 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the 1923 destination, C<nitems> is the number of items, and C<type> is the type. 1924 1925 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type 1926 1927 Like C<Zero> but returns dest. Useful 1928 for encouraging compilers to tail-call 1929 optimise. 1930 1931 =for apidoc Am|void|StructCopy|type *src|type *dest|type 1932 This is an architecture-independent macro to copy one structure to another. 1933 1934 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte 1935 1936 Fill up memory with a byte pattern (a byte repeated over and over 1937 again) that hopefully catches attempts to access uninitialized memory. 1938 1939 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type 1940 1941 PoisonWith(0xAB) for catching access to allocated but uninitialized memory. 1942 1943 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type 1944 1945 PoisonWith(0xEF) for catching access to freed memory. 1946 1947 =for apidoc Am|void|Poison|void* dest|int nitems|type 1948 1949 PoisonWith(0xEF) for catching access to freed memory. 1950 1951 =cut */ 1952 1953 /* Maintained for backwards-compatibility only. Use newSV() instead. */ 1954 #ifndef PERL_CORE 1955 #define NEWSV(x,len) newSV(len) 1956 #endif 1957 1958 #define MEM_SIZE_MAX ((MEM_SIZE)~0) 1959 1960 1961 #ifdef PERL_MALLOC_WRAP 1962 1963 /* This expression will be constant-folded at compile time. It checks 1964 * whether or not the type of the count n is so small (e.g. U8 or U16, or 1965 * U32 on 64-bit systems) that there's no way a wrap-around could occur. 1966 * As well as avoiding the need for a run-time check in some cases, it's 1967 * designed to avoid compiler warnings like: 1968 * comparison is always false due to limited range of data type 1969 * It's mathematically equivalent to 1970 * max(n) * sizeof(t) > MEM_SIZE_MAX 1971 */ 1972 1973 # define _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) \ 1974 (8 * sizeof(n) + sizeof(t) > sizeof(MEM_SIZE)) 1975 1976 /* This is written in a slightly odd way to avoid various spurious 1977 * compiler warnings. We *want* to write the expression as 1978 * _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) && (n > C) 1979 * (for some compile-time constant C), but even when the LHS 1980 * constant-folds to false at compile-time, g++ insists on emitting 1981 * warnings about the RHS (e.g. "comparison is always false"), so instead 1982 * we write it as 1983 * 1984 * (cond ? n : X) > C 1985 * 1986 * where X is a constant with X > C always false. Choosing a value for X 1987 * is tricky. If 0, some compilers will complain about 0 > C always being 1988 * false; if 1, Coverity complains when n happens to be the constant value 1989 * '1', that cond ? 1 : 1 has the same value on both branches; so use C 1990 * for X and hope that nothing else whines. 1991 */ 1992 1993 # define _MEM_WRAP_WILL_WRAP(n,t) \ 1994 ((_MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) ? (MEM_SIZE)(n) : \ 1995 MEM_SIZE_MAX/sizeof(t)) > MEM_SIZE_MAX/sizeof(t)) 1996 1997 # define MEM_WRAP_CHECK(n,t) \ 1998 (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \ 1999 && (croak_memory_wrap(),0)) 2000 2001 # define MEM_WRAP_CHECK_1(n,t,a) \ 2002 (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \ 2003 && (Perl_croak_nocontext("%s",(a)),0)) 2004 2005 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t), 2006 2007 #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1))) 2008 #else 2009 2010 #define MEM_WRAP_CHECK(n,t) 2011 #define MEM_WRAP_CHECK_1(n,t,a) 2012 #define MEM_WRAP_CHECK_2(n,t,a,b) 2013 #define MEM_WRAP_CHECK_(n,t) 2014 2015 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1))) 2016 2017 #endif 2018 2019 #ifdef PERL_MEM_LOG 2020 /* 2021 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s 2022 * go through functions, which are handy for debugging breakpoints, but 2023 * which more importantly get the immediate calling environment (file and 2024 * line number, and C function name if available) passed in. This info can 2025 * then be used for logging the calls, for which one gets a sample 2026 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined. 2027 * 2028 * Known problems: 2029 * - not all memory allocs get logged, only those 2030 * that go through Newx() and derivatives (while all 2031 * Safefrees do get logged) 2032 * - __FILE__ and __LINE__ do not work everywhere 2033 * - __func__ or __FUNCTION__ even less so 2034 * - I think more goes on after the perlio frees but 2035 * the thing is that STDERR gets closed (as do all 2036 * the file descriptors) 2037 * - no deeper calling stack than the caller of the Newx() 2038 * or the kind, but do I look like a C reflection/introspection 2039 * utility to you? 2040 * - the function prototypes for the logging functions 2041 * probably should maybe be somewhere else than handy.h 2042 * - one could consider inlining (macrofying) the logging 2043 * for speed, but I am too lazy 2044 * - one could imagine recording the allocations in a hash, 2045 * (keyed by the allocation address?), and maintain that 2046 * through reallocs and frees, but how to do that without 2047 * any News() happening...? 2048 * - lots of -Ddefines to get useful/controllable output 2049 * - lots of ENV reads 2050 */ 2051 2052 # ifdef PERL_CORE 2053 # ifndef PERL_MEM_LOG_NOIMPL 2054 enum mem_log_type { 2055 MLT_ALLOC, 2056 MLT_REALLOC, 2057 MLT_FREE, 2058 MLT_NEW_SV, 2059 MLT_DEL_SV 2060 }; 2061 # endif 2062 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */ 2063 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); 2064 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); 2065 # endif 2066 # endif 2067 2068 #endif 2069 2070 #ifdef PERL_MEM_LOG 2071 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__) 2072 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__) 2073 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__) 2074 #endif 2075 2076 #ifndef MEM_LOG_ALLOC 2077 #define MEM_LOG_ALLOC(n,t,a) (a) 2078 #endif 2079 #ifndef MEM_LOG_REALLOC 2080 #define MEM_LOG_REALLOC(n,t,v,a) (a) 2081 #endif 2082 #ifndef MEM_LOG_FREE 2083 #define MEM_LOG_FREE(a) (a) 2084 #endif 2085 2086 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) 2087 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) 2088 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t))))) 2089 2090 #ifndef PERL_CORE 2091 /* pre 5.9.x compatibility */ 2092 #define New(x,v,n,t) Newx(v,n,t) 2093 #define Newc(x,v,n,t,c) Newxc(v,n,t,c) 2094 #define Newz(x,v,n,t) Newxz(v,n,t) 2095 #endif 2096 2097 #define Renew(v,n,t) \ 2098 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) 2099 #define Renewc(v,n,t,c) \ 2100 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) 2101 2102 #ifdef PERL_POISON 2103 #define Safefree(d) \ 2104 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0) 2105 #else 2106 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d))) 2107 #endif 2108 2109 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) 2110 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) 2111 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t))) 2112 2113 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) 2114 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) 2115 #ifdef HAS_MEMSET 2116 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t))) 2117 #else 2118 /* Using bzero(), which returns void. */ 2119 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d) 2120 #endif 2121 2122 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t))) 2123 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB) 2124 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF) 2125 #define Poison(d,n,t) PoisonFree(d,n,t) 2126 2127 #ifdef PERL_POISON 2128 # define PERL_POISON_EXPR(x) x 2129 #else 2130 # define PERL_POISON_EXPR(x) 2131 #endif 2132 2133 #ifdef USE_STRUCT_COPY 2134 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s))) 2135 #else 2136 #define StructCopy(s,d,t) Copy(s,d,1,t) 2137 #endif 2138 2139 /* C_ARRAY_LENGTH is the number of elements in the C array (so you 2140 * want your zero-based indices to be less than but not equal to). 2141 * 2142 * C_ARRAY_END is one past the last: half-open/half-closed range, 2143 * not last-inclusive range. */ 2144 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0])) 2145 #define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a)) 2146 2147 #ifdef NEED_VA_COPY 2148 # ifdef va_copy 2149 # define Perl_va_copy(s, d) va_copy(d, s) 2150 # else 2151 # if defined(__va_copy) 2152 # define Perl_va_copy(s, d) __va_copy(d, s) 2153 # else 2154 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list) 2155 # endif 2156 # endif 2157 #endif 2158 2159 /* convenience debug macros */ 2160 #ifdef USE_ITHREADS 2161 #define pTHX_FORMAT "Perl interpreter: 0x%p" 2162 #define pTHX__FORMAT ", Perl interpreter: 0x%p" 2163 #define pTHX_VALUE_ (void *)my_perl, 2164 #define pTHX_VALUE (void *)my_perl 2165 #define pTHX__VALUE_ ,(void *)my_perl, 2166 #define pTHX__VALUE ,(void *)my_perl 2167 #else 2168 #define pTHX_FORMAT 2169 #define pTHX__FORMAT 2170 #define pTHX_VALUE_ 2171 #define pTHX_VALUE 2172 #define pTHX__VALUE_ 2173 #define pTHX__VALUE 2174 #endif /* USE_ITHREADS */ 2175 2176 /* Perl_deprecate was not part of the public API, and did not have a deprecate() 2177 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor 2178 CPAN::Unpack show any users outside the core. */ 2179 #ifdef PERL_CORE 2180 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \ 2181 "Use of " s " is deprecated") 2182 #endif 2183 2184 /* Internal macros to deal with gids and uids */ 2185 #ifdef PERL_CORE 2186 2187 # if Uid_t_size > IVSIZE 2188 # define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid)) 2189 # define SvUID(sv) SvNV(sv) 2190 # else 2191 # if Uid_t_sign <= 0 2192 # define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid)) 2193 # define SvUID(sv) SvIV(sv) 2194 # else 2195 # define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid)) 2196 # define SvUID(sv) SvUV(sv) 2197 # endif 2198 # endif /* Uid_t_size */ 2199 2200 # if Gid_t_size > IVSIZE 2201 # define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid)) 2202 # define SvGID(sv) SvNV(sv) 2203 # else 2204 # if Gid_t_sign <= 0 2205 # define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid)) 2206 # define SvGID(sv) SvIV(sv) 2207 # else 2208 # define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid)) 2209 # define SvGID(sv) SvUV(sv) 2210 # endif 2211 # endif /* Gid_t_size */ 2212 2213 #endif 2214 2215 #endif /* HANDY_H */ 2216 2217 /* 2218 * ex: set ts=8 sts=4 sw=4 et: 2219 */ 2220